Practical guide
AI fluency for people, learning and organisation teams
Plan capability-development exercises without confusing course completion, tool activity and demonstrated workplace performance.
People and learning teams can use AI-enabled exercises to investigate development needs, design practice and interpret task evidence. This original guide separates learning participation from observed performance and organisational decisions. It does not position an assessment as an employment decision tool or claim that one exercise can establish a person's suitability for a role.
Name the development question
Decide whether the task concerns knowledge, participation, applied work or organisational conditions.
A fictional organisation has completed an introductory AI course but wants to know what support employees need next. Course attendance answers who participated. A knowledge quiz asks what concepts someone can recall or explain. A realistic task asks how a person works with supplied materials under particular conditions.
Write the question before choosing an instrument. Do not use a convenient attendance report as a substitute for evidence of applied work. Equally, do not dismiss participation data when the actual question is access to training. Different evidence serves different decisions.
Map practice to the work context
Select a task that resembles the intended learning outcome without importing unnecessary personal information.
For the fictional programme, one group drafts a supplier brief and another prepares a customer handoff. Both may need source checking, but their artifacts and consequences differ. Specify the relevant materials, constraints and expected output for each. Use synthetic or appropriately approved inputs.
Ask AI to help draft practice variations, then check whether they still require the intended behavior. An exercise designed to test verification should contain something verifiable and a realistic reason to check it. A generic writing assignment cannot reveal every capability simply because AI was available.
Keep observation and interpretation separate
A development report should say what was visible before assigning meaning to it.
The participant may correct an unsupported claim, cite a source or revise a recommendation after new information. Those actions can support a task-specific discussion. A missing log entry, however, may reflect limited visibility rather than absence of the behavior. Record that limitation instead of filling it with a negative judgment.
Use a small evidence note: observed action, relevant artifact, interpretation and uncertainty. Avoid personality labels or broad statements about future performance. A learning conversation should remain connected to work the participant can inspect and practice differently.
Choose support at the right level
A difficulty may come from an individual learning need, a task design problem or an organisational constraint.
In the fictional programme, several participants cannot verify a claim because the necessary source is absent. More training on diligence would not repair the exercise. If the source is available but its relevance is unclear, instruction on evidence selection may help. If approved tools cannot access it, investigate the workflow.
Ask the programme owner to distinguish these possibilities before assigning remediation. AI can help organize observations, but it should not automatically decide who needs what intervention from incomplete evidence. Preserve a route for contextual review and correction.
Evaluate transfer with a different task
A follow-up should test whether the practiced behavior appears beyond the exact exercise used for instruction.
Change the materials and business context while keeping the target behavior recognizable. For example, move from checking a supplier claim to checking a project-status assertion. Review whether the participant identifies the relevant source and preserves uncertainty. Do not treat memorizing the original correction as evidence of broader transfer.
Keep the comparison conditions visible, including differences in task difficulty and support. Use the result to refine the learning plan rather than to make an automated employment recommendation. The learning and development manager guide describes a specific transfer-check design; this function guide maps the wider set of evidence and programme decisions.
Sources and scope
NIST addresses AI risk management. Skills England describes workplace AI foundations.
These sources provide background, not endorsement of this exercise. The worked example and suggested review method are original illustrative guidance. They are not customer results, validated benchmarks or evidence of a particular product capability. Adapt the exercise to the task and use qualified review where consequences require it.
Sources: [1] [2]
Sources
- 1.AI RMF Core · NIST
- 2.AI foundation skills for work benchmark · Skills England